It is now recognized in many domains that content-based image retrieval fro
m a database of images cannot be carried out by using completely automated
approaches. One such domain is medical radiology for which the clinically u
seful information in an image typically consists of gray level variations i
n highly localized regions of the image. Currently, it is not possible to e
xtract these regions by automatic image segmentation techniques. To address
this problem, we have implemented a human-in-the-loop (a physician-in-the-
loop, more specifically) approach in which the human delineates the patholo
gy bearing regions (PBR) and a set of anatomical landmarks in the image whe
n the image is entered into the database. To the regions thus marked, our a
pproach applies low-level computer vision and image processing algorithms t
o extract attributes related to the variations in gray scale, texture, shap
e, etc. In addition, the system records attributes that capture relational
information such as the position of a PER with respect to certain anatomica
l landmarks. An overall multidimensional index is assigned to each image ba
sed on these attribute values. (C) 1999 Academic Press.